On-Line Trajectory-Based Linearisation of Neural Models for a Computationally Efficient Predictive Control Algorithm

2012 ◽  
pp. 126-134 ◽  
Author(s):  
Maciej Ławryńczuk
Keyword(s):  
Predictive Control ◽  
Control Algorithm ◽  
Computationally Efficient ◽  
Neural Models ◽  
On Line ◽  
Line Trajectory

scholarly journals Efficient Nonlinear Predictive Control Based on Structured Neural Models

2009 ◽  
Vol 19 (2) ◽  
pp. 233-246 ◽  
Author(s):  
Maciej Ławryńczuk
Keyword(s):  
Predictive Control ◽  
Neural Model ◽  
External Input ◽  
Computationally Efficient ◽  
Neural Models ◽  
Loop Control ◽  
Nonlinear Predictive Control ◽  
On Line ◽  
Auto Regressive ◽  
Narx Models

Efficient Nonlinear Predictive Control Based on Structured Neural ModelsThis paper describes structured neural models and a computationally efficient (suboptimal) nonlinear Model Predictive Control (MPC) algorithm based on such models. The structured neural model has the ability to make future predictions of the process without being used recursively. Thanks to the nature of the model, the prediction error is not propagated. This is particularly important in the case of noise and underparameterisation. Structured models have much better long-range prediction accuracy than the corresponding classical Nonlinear Auto Regressive with eXternal input (NARX) models. The described suboptimal MPC algorithm needs solving on-line only a quadratic programming problem. Nevertheless, it gives closed-loop control performance similar to that obtained in fully-fledged nonlinear MPC, which hinges on online nonconvex optimisation. In order to demonstrate the advantages of structured models as well as the accuracy of the suboptimal MPC algorithm, a polymerisation reactor is studied.

scholarly journals Rollover Prevention System Dedicated to ATVs on Natural Ground

2012 ◽  
Vol 162 ◽  
pp. 505-514 ◽  
Author(s):  
Mathieu Richier ◽  
Roland Lenain ◽  
Benoit Thuilot ◽  
Christophe Debain
Keyword(s):  
Predictive Control ◽  
Control Algorithm ◽  
Load Transfer ◽  
Low Cost ◽  
Lateral Load ◽  
Dynamical Model ◽  
Rollover Prevention ◽  
Perception System ◽  
Prevention System ◽  
On Line

In this paper, an algorithm dedicated to light ATVs, which estimates and anticipates the rollover, is proposed. It is based on the on-line estimation of the Lateral Load Transfer (LLT), allowing the evaluation of dynamic instabilities. The LLT is computed thanks to a dynamical model split into two 2D projections. Relying on this representation and a low cost perception system, an observer is proposed to estimate on-line the terrain properties (grip conditions and slope), then allowing to deduce accurately the risk of instability. Associated to a predictive control algorithm, based on the extrapolation of riders action, the risk can be anticipated, enabling to warn the pilot and to consider the implementation of active actions.

Generalized predictive control of a robotic manipulator with hydraulic actuators

Robotica ◽  
1992 ◽  
Vol 10 (5) ◽  
pp. 447-459 ◽  
Author(s):  
A. Kotzev ◽  
D. B. Cherchas ◽  
P. D. Lawrence ◽  
N. Sepehri
Keyword(s):  
Adaptive Control ◽  
Predictive Control ◽  
Control Algorithm ◽  
Generalized Predictive Control ◽  
Maximum Output ◽  
Hydraulic Actuators ◽  
Actuation System ◽  
Hydraulically Actuated ◽  
On Line ◽  
Adaptive Control Algorithm

SUMMARYThis paper presents some aspects of the behavior of hydraulically actuated heavy duty manipulators. This category of manipulators is used extensively in large resource based industries and any improvement in efficiency may result in major financial benefits. In this paper an adaptive control algorithm is used for a two rigid link manipulator driven by hydraulic actuators. The dynamic model of the manipulator is derived as well as the models of the hydraulic actuators including compliance, dead time and full dynamics of the servo valves. An adaptive control algorithm is considered since changes occur on-line in the system's parameters. The adaptive algorithm used is Generalized Predictive Control (GPC). The GPC uses a controlled autoregressive integrated moving average (CARIMA) type model and a cost function that minimizes a predicted future output error and future weighted control inputs to the plant, resulting in a sequence of future control increments. The procedure, in this work, does not separate the hydraulic actuator and the link dynamics into separate sub-systems, but controls them as one system. The changes in the system's parameters due to the hydraulics or the link dynamics can be estimated and the coefficients of the model adjusted without the necessity of identifying the exact cause of the changes.It was found in this work that the variations of the GPC control horizon can lead to faster response during transients and significantly reduced overshoot in the nonlinear hydraulic actuation system. An on-line change of the maximum output horizon is also introduced.This work shows the analysis and results of a two link manipulator with hydraulic actuators. It can be implemented on any hydraulically actuated manipulator with any number of links and actuators.Numerical simulations are performed on a Vax 3200 computer and the results are presented.

scholarly journals Nonlinear predictive control based on neural multi-models

2010 ◽  
Vol 20 (1) ◽  
pp. 7-21 ◽  
Author(s):  
Maciej Ławryńczuk ◽  
Piotr Tatjewski
Keyword(s):  
Predictive Control ◽  
Prediction Error ◽  
Control Policy ◽  
External Input ◽  
Multi Layer Perceptron ◽  
Computationally Efficient ◽  
Nonlinear Predictive Control ◽  
Nonlinear Optimisation ◽  
On Line ◽  
Auto Regressive

Nonlinear predictive control based on neural multi-modelsThis paper discusses neural multi-models based on Multi Layer Perceptron (MLP) networks and a computationally efficient nonlinear Model Predictive Control (MPC) algorithm which uses such models. Thanks to the nature of the model it calculates future predictions without using previous predictions. This means that, unlike the classical Nonlinear Auto Regressive with eXternal input (NARX) model, the multi-model is not used recurrently in MPC, and the prediction error is not propagated. In order to avoid nonlinear optimisation, in the discussed suboptimal MPC algorithm the neural multi-model is linearised on-line and, as a result, the future control policy is found by solving of a quadratic programming problem.

Sign in / Sign up

Export Citation Format

Share Document